黄仁勋希望世界停止对人工智能的担忧，转而开始利用它进行构建。在英伟达于柏林举办的GTC活动上，他驳斥了人类灭绝的预言，为AI的就业前景辩护，并推介一款售价4000美元的电脑作为家庭机器人。这位英伟达掌门人的盲目乐观在问答环节中偶尔流露出一丝恼怒。
Jensen Huang wants the world to stop fretting about AI and start building with it. At Nvidia's GTC event in Berlin, he dismissed predictions of human extinction, defended AI's job prospects, and pitched a $4,000 computer as the family robot. The Nvidia chief's relentless optimism occasionally gave way to irritation during a question-and-answer session. "If we don't describe a future worthy to be part of, then we're not going to lead people there," he said. "We have to describe a future worthy of creating. We have to describe a future not of fear, but of optimism. Who wants to create a future that's worse than today's?
“如果我们不描绘一个值得参与的未来，就无法带领人们走向那里，”他说。“我们必须描绘一个值得创造的未来。我们必须描绘一个没有恐惧、充满乐观的未来。谁想创造一个比今天更糟糕的未来？我们不必费力去创造一个比今天更糟糕的未来，只要听之任之就行，不是吗？”
We don't have to work hard to create a future worse than today. We just let it happen, isn't that right?" Huang said that we still have to ask the hard questions about safety and jobs, but "it's important, of course, that we don't make it up. It is not helpful in describing an optimistic future when you say that it will be the end of the human civilization. First of all, it's not true.
黄仁勋表示，我们仍然必须就安全和就业提出尖锐问题，但“重要的是，当然，我们不要凭空捏造。当你说那将是人类文明的终结时，这对描绘一个乐观的未来没有帮助。首先，那不是真的。那是编造出来的。我知道那能博眼球，”他说，将此类报道斥为标题党。
It's made up. I know it gets headlines," he said, dismissing such stories as clickbait. When asked if AI really makes people more productive, he explained that there are two ways to think about that: either it helps you do the same old thing more efficiently, or it helps you do things you couldn't before.
当被问及AI是否真的让人类更有生产力时，他解释说，对此有两种思考方式：要么它帮你更高效地做同样的旧事，要么它帮你做以前做不到的事。
"Most of engineering, most of design, has to do with exploring options. Try, experiment, try and experiment, iterate, experiment, iterate," he said. "What's the best combination?
“大多数工程、大多数设计，都与探索选项有关。尝试、实验、尝试再实验、迭代、实验、再迭代，”他说。“什么是最佳组合？而你只有有限的时间去尝试不同版本。现在我有了AI去探索那个宇宙。那就是生产力。而在我们公司内部，我们现在每天都在这样做。我们正在探索符合我们关心的约束条件、物理或现实约束条件的设计选项宇宙。”黄仁勋指出，AI数据中心热潮催生了建筑和基础设施工作。
And you can only have so much time to try different versions. Now I have AI to go off and explore that universe. That is productivity. And inside our company, we do this every single day now. We're exploring the universe of design options that are within the constraints that we care about, that are physical or realistic constraints." Huang pointed to construction and infrastructure work generated by the AI datacenter boom.
“土木工人、水管工、电工、网络工程师，所有这些令人难以置信的工作岗位都在美国被创造出来。我认为这已经创造了100万个工作岗位，再看看在这之上，有多少初创企业正在被创立，”他宣称道。“这就是生产力，提高GDP，增加我们的增长，这也导致我们变得更忙碌。但归根结底，AI创造工作，希望不会夺走工作。”
"Construction workers, plumbers, electricians, networking engineers, all of those incredible jobs are being created in the United States. I think it already has created a million jobs, and then the layer above that, look at all the startups that are being created," he claimed. "That's productivity, increasing the GDP, increasing our growth, which also results in increasing how busy we are. But that's the end. AI makes jobs, and hopefully doesn't take jobs."
被问及人们已经因AI而失业——全球金融服务集团巴克莱银行今天上午刚宣布裁减5000名员工并更多依赖AI——黄仁勋辩称，技术变革一直以来都在取代一些工作的同时创造另一些工作。“现在，有了AI，当然，一些工作不像以前那样必要了，或者可能不像以前那样有趣了。而一些新工作、新公司和新业务变得必要了。这是世界的本质，”他说道。
Asked about people already losing work to AI - global financial services group Barclays Bank said only this morning it is chopping 5,000 workers and relying more on AI - Huang argued that technological change has always displaced some jobs while creating others. "Now, with AI, of course, some jobs are not necessary as much as before, or maybe not as interesting as before. And some new jobs and new companies and new businesses are necessary. It's the nature of the world," he stated. On safety, Huang pointed to Nvidia's OpenShell sandbox technology, designed to restrict agents to the resources needed for a task. "If you're making an omelette, you don't need a hammer," he said.
在安全问题上，黄仁勋指出了英伟达的OpenShell沙盒技术，该技术旨在将智能体限制在完成任务所需的资源范围内。“如果你在做煎蛋卷，你不需要一把锤子，”他说。
But he got hot under the collar when asked about slowing down the pace of AI development while the risks were evaluated, in light of incidents such as AI agents from Anthropic and OpenAI breaking into other computer systems. "You would favor pacing the frontier? I favor them doing whatever they need to do to not bring harm to society," Huang said.
但当被问及在评估风险时是否应放缓AI发展步伐，特别是考虑到Anthropic和OpenAI的AI智能体闯入其他计算机系统等事件时，他勃然大怒。“你主张放缓前沿发展？我主张他们做任何必要的事以避免给社会带来危害，”黄仁勋说。“先生，如果你认为你不该出门，因为你穿得不得体，或者以某种方式会给社会带来危害，能不能请你别出来？”他对提问者说道。
"You, sir, if you believe that you should not leave your house because you're inappropriately dressed, or somehow you're going to bring harm to society, could you please not come out?" he told the questioner. "The message from the labs is that there's a collective action problem, and that some minimum standards in coordination would be good here. That's a different question. If you know something is already delivering harm, you got to stop running," he added. Nvidia has a commercial interest in keeping the AI build-out moving.
“实验室传递的信息是存在集体行动问题，在此进行一些最低标准的协调会很好。那是另一个问题。如果你知道某事已经在造成危害，你就必须停止奔跑，”他补充道。
It reported $215.9 billion in revenue for fiscal 2026 and $120 billion in net profit. In the first half of its fiscal 2027, those net profits were already close to the prior year's . The opportunity isn't limited to just chips either.
英伟达在推动人工智能技术的发展方面具有明显的商业利益。该公司在2026财年的收入达到了2159亿美元，净利润为120亿美元；而在2027财年的上半年，其净利润已经接近上一财年的水平。英伟达的盈利机会并不仅限于芯片业务。黄（Huang）表示，英伟达软件套件的年许可费用约为每块GPU 2000美元。这位首席执行官还展示了英伟达产品的家庭应用场景：当被问及普通家庭何时能够运行大型人工智能模型时，他举出了一台DGX Spark系统作为例子。
Huang put the price of licensing Nvidia's software stack at about $2,000 per GPU annually. The CEO's domestic vision was also on show. Asked when households could run large AI models, he held up a DGX Spark system. "This computer has a chip called N1X inside, and N1X has 128 gigabytes of memory… It's $4,000. You're right. It's quite expensive. However, that computer was designed to sit on the network in your house, and it would have agents that are used by all of your family members. So it's kind of like a TV. It's sitting in your house. It's shared by everyone, and everybody's agents will be running on DGX Spark," he explained.
“这台电脑内部配备了N1X芯片，该芯片拥有128GB的内存，售价为4000美元。确实很贵，但它被设计用于家庭网络环境中，所有家庭成员都可以使用其中的AI应用程序。它就像一台电视一样，被放在家里供大家共享使用。”黄举例说，他家中的AI系统曾研究过柏林最好的炸肉排，并生成了一份包含餐厅图片和菜单的宣传册。
As an example, Huang said his home agent had researched Berlin's best schnitzel and produced a brochure with restaurant pictures and menus. "So that agent, you know, that computer will become essentially your family's robot. It has no arms and legs, but it's an agent, so it's kind of like a robot. And this one is just for your home. So that's kind of how I see the future. This is pretty affordable.
“这种AI系统本质上就是你的家庭‘机器人’——虽然没有四肢，但它确实具备智能处理能力，因此可以被视为一种‘机器人’。这样的产品其实非常实用且价格合理（前提是你的净资产在1800亿到1900亿美元之间）；不过，对于那些在生活成本危机中难以支付账单的普通家庭来说，可能仍然难以负担。”
This is a personal computer." Affordable if your net worth is around $180-$190 billion. Maybe not to the average household struggling to pay bills in a cost of living crisis. ®